摘要
目的使用R语言分析模型对结核病发病率数据进行演算与分析,探索较佳的分析模型。方法使用重庆市1993—2007年结核病发病登记数作为建模数据,采用R语言中的神经网络模型和灰色模型分别预测2008年与2009年的发病人数,并与其对应年份的真实发病人数进行比较。结果采用神经网络模型预测结果误差分别为4.75%、0.83%,灰色模型误差分别为27.08%、49.84%。结论神经网络模型为较佳分析模型,预测结果对实际工作的量化和资源调配有重要的指导意义。
Objective To compare two predicting models for tuberculosis (TB) incidence using R language. Methods Recorded numbers of TB patients in Chongqing between 1993 and 2007 were used as modeling data. Neuro - network model and gray model programmed with R language were used to predict the TB incidence in 2008 and 2009, respectively. These pre- dicted TB incidences were compared with the real TB incidences in 2008 and 2009. Results The predicting errors of neuro - net- work model were 4.75% for 2008 and 0. 83% for 2009, while those of gray model were 27.08% for 2008 and 49. 84% for 2009. Conclusion The neuro - network model is more accurate in predicting TB incidence than the gray model.
出处
《中国全科医学》
CAS
CSCD
北大核心
2013年第13期1498-1501,共4页
Chinese General Practice
基金
国家自然科学基金(30872160)
重庆市科委自然科学基金(CSTC2009BB5415)
关键词
结核病
R语言
预测模型
Tuberculosis
R language
Predicting model